Philosophy

http://journals.cambridge.org/PHI

Additional services for Philosophy:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

On Saying and Showing

A. W. Moore

Philosophy / Volume 62 / Issue 242 / October 1987, pp 473 - 497

DOI: 10.1017/S0031819100039048, Published online: 30 January 2009

Link to this article: http://journals.cambridge.org/abstract_S0031819100039048

How to cite this article:

A. W. Moore (1987). On Saying and Showing. Philosophy, 62, pp 473-497 doi:10.1017/
S0031819100039048

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/PHI, IP address: 130.88.20.147 on 10 Nov 2015

I

t
On Saying and Showing
A. W. MOORE

There is not, and may there never be, any treatise by me . . .on these
things, for the subject is not communicable in words, as other
sciences are. Rather is it that, after long association in the business
itself and a shared life, a light is lit in the soul, kindled, as it were, by a
leaping flame, and thenceforward feeds itself (Plato).

1. The Saying/Showing Distinction in the Tractatus. Two

Examples

A linchpin of Wittgenstein's Tractatus1 is his distinction between what

can be said and what can be shown. The distinction first seriously
emerges in the early 4's in connection with the idea that a proposition
shows, but does not say, what its sense is (4.022ff.).2 What can be
shown, we are told, cannot be said (4.1212). By the end of the work the
catalogue of what can be shown has become large and multifarious. It
includes: the logical form of reality (4.121); the logical relations
between propositions (4.1211, 6.1201 and 6.1221); the limit of empiri-
cal reality (5.5561); the truth in solipsism, that is what the solipsist
means (5.62); and the mystical (6.522). Recent work bearing on the
distinction, and sympathetic to it, has pointed to an even broader
conception of what can be shown. For example, it has been suggested
that one lesson of Wittgenstein's later work, where the terrain is
notoriously very different from that of the Tractatus, and where the
distinction is certainly not explicitly retained, is that there are certain
things which can be shown though they cannot properly be put into
words.3 My aim in this essay is to probe the very idea of a saying/show-

1
Ludwig Wittgenstein, Tractatus Logico-Philosophicus, trans. D. F.
Pears and B. F. McGuiness (London: Routledge and Kegan Paul, 1961).
2
But the first allusion to the distinction is in 2.171—2.172.
3
See e.g. Bernard Williams, 'Wittgenstein and Idealism', reprinted in
Moral Luck (Cambridge University Press, 1981), 163, and Jonathan Lear,
'The Disappearing "We"', in Proceedings of the Aristotelian Society, Sup-
plementary Volume 58, (1984), 242. I have tried to develop this line of
thought in 'Transcendental Idealism in Wittgenstein, and Theories of
Meaning', in Philosophical Quarterly 35 (1985).

Philosophy 62 1987 473

I
A. W. Moore

ing distinction along Tractarian lines. I shall attempt to say what it

consists in, what philosophical work it has to do, and how one might be
justified in drawing such a distinction.
Consider the two following examples. The first is inspired by the
Tractatus (see 5.633—5.6331). It concerns somebody's visual field,
where this is to be thought of as a three-dimensional portion of public
space: a more or less complete description is produced of what is in it,
from the point of view of the person, though with no explicit reference
to anything outside the field. However complete the description may
be, it cannot represent the fact that everything of which it treats is seen
from a particular point at the edge of the field. For this is not itself a fact
exclusively about what is in the field. Yet there is a sense in which this
fact will be manifest in the form that the description takes. (The
description will use terms like 'left' and 'right'.) So there is something
which cannot be said here but which is shown by what can be said.
The second example4 concerns a limited language with a co-ordinate
system by means of which things are located on the surface of the earth.
Every sentence from the language is of the form 'X is at (i, j ) ' . It is clear
that this language will be fully comprehensible only to somebody who
can identify some point on the grid. Yet none of its sentences can be
used to say where any point on the grid is. At most they can be used to
say what is at any given point. What cannot be said here must again be
in some sense manifest to, or shown to, anybody with a full understand-
ing of what can be said.5
Is this the kind of thing which Wittgenstein is alluding to in the
Tractatus} In at least one crucial respect, it is not. Wittgenstein wants
to claim that there are things which can be shown though they cannot be
said in any possible language, whereas each of the examples furnishes
us with a notion of unsayability which is relative to certain limited
linguistic resources. In each case what is shown can also be said, given
appropriate enriching of those resources. The examples nevertheless
furnish a useful model of Wittgenstein's distinction, and they provide
some clues as to what unifying theme might link his inventory of what
can be shown. In each of the examples what is showable is some feature
of a framework, and what is sayable is how that framework is filled in.
The frameworks are the outer limits of the visual field and the co-
ordinate system respectively. Each of these determines a certain
bounded whole, a space, and it is the parts of this space which are

4
This example is derived from an unpublished lecture course given by
Ron Williams at Colorado State University.
5
I believe that there are also some useful examples to be gleaned from
mathematics; see e.g. my 'Set Theory, Skolem's Paradox and the Trac-
tatus', in Analysis 45 (1985).

474
* On Saying and Showing

t
described whenever anything is said (within the prescribed linguistic
limits); but nothing that can be said (within those limits) serves to
describe the whole itself, as a whole—as part of some larger space. By
contrast, what can be shown does concern the whole as a whole, related
to something outside it, namely a subject. In the first instance it is
shown that the contents of the whole are seen in a certain way by the
subject. In the second instance it is shown that the contents of the whole
are to be located in a certain way by the subject.
Now the contingencies of any relationship between the whole and
what lies outside it are irrelevant to the description of its parts. For the
purpose of such a description, the whole, or rather its framework,
counts as rigid and fixed. So, relative to that description, what can be
said is how things are within the framework and what can be shown,
bearing as it does on the framework itself, explains how they must be.
The distinction between what can be said and what can be shown
therefore relates in a fundamental way to the familiar philosophical
distinction between what is actually the case and what is necessarily the
case. This is indeed reflected in the Tractatus, where it is claimed that
the only necessity is logical necessity (6.37). For elsewhere it is clear
that Wittgenstein intends an intimate connection between logic (logical
form and structure) and what can be shown (for example, 4.121).6
In general, this model does give us some sense of how the various
doctrines advanced in the Tractatus concerning the saying/showing
distinction might be held together. The relevant whole in the Tractatus
is the world, all that is the case. Its shape and its form, when it is
conceived as a whole (a limited whole), are what can be shown. And
through the ethical and mystical elements of the work these are related
to the metaphysical subject, which lies outside the world, as its limit.

2. Some Preliminaries
If these ideas are to be sharpened, however, something still needs to be
said about the special way in which the word 'show' is being used and,
more particularly, about the grammar of the saying/showing distinc-
tion. Otherwise the claim that certain things can be shown though they
cannot be said will be too easily understood as a triviality which falls far
short of Wittgenstein's thesis. (I can show you a chair but I cannot say
it.) It is not just that 'things' is too imprecise. Any narrower sortal
which does not issue in the same kind of trivial interpretation will result
in a claim which Wittgenstein would want to deny—for example, that
certain truths can be shown but not said, or, worse still, that what

6
5.634 is also relevant here.

475
A. W. Moore

certain states of affairs are like can be shown but not said (see, for
example, 4.063). Concomitant with the problems which arise here is a
question which has received surprisingly little attention and with
respect to which the Tractatus offers no help: what is the genus of
which saying and showing are species? Of course, there is a sense in
which this is a bad question, born precisely of an overly crude construal
of the claim that certain things can be shown but not said. We are less
inclined to pose the question when we return to the text. For although
Wittgenstein speaks of propositions as both saying and showing certain
things, he also suggests that the real contrast is between what we
(language-users) say by means of propositions and what shows itself, or
makes itself manifest (for example, 4.121 and 6.124).7 Still, to the
extent that a division is being made here, we are bound to want some
sense of the territory which is being divided.
In the light of this, it is appropriate (though not particularly Trac-
tarian) to focus on individual states of enlightenment—to see the
distinction as being, primarily, between knowing something that can be
said (as it were, having the answer to some question (see 6.5)) and being
shown something, where this is construed as being in a state of
enlightenment which is not of that kind. The claim that certain things
can be shown though they cannot be said is then tantamount to the
claim that we can be in a state of enlightenment though we are unable to
give voice to it (we are unable to put it into words). If we view things in
this way, then we are no longer likely to be vexed by the question of
what the common object of saying and showing is or by the question of
what their genus is. We nevertheless do have something to say about
what are being distinguished, namely states of enlightenment. More-
over, by turning our attention away from what is shown to those who
are shown it, we help to allay the natural worry that much of what we
shall want to say in this connection is precisely what, according to the
doctrine being discussed, cannot be said. For whatever we can or
cannot say about what is shown, we are certainly able to say plenty
about people and about what it is for them to be in states of enlighten-
ment. Thus we may be able to say a lot about showing without prejudic-
ing any issues, and in particular without abrogating the very concept
under discussion. Were this not the case, there would be legitimate
room for suspicion about whether it was a bonafide concept at all.

3. Basic Justification for Drawing such a Distinction

Let us focus, then, on ourselves as knowing subjects. The following are
some elemental theses about us in that capacity and they would be
7
Max Black makes this point in A Companion to Wittgenstein's 'Trac-
tatus' (Cambridge University Press, 1964), 190.

476
 On Saying and Showing

regarded by many as axiomatic: we are cast into a world which is not of

our own making (which is independent of us); it is this world of which
we have knowledge (if we have any knowledge at all); and, because our
epistemic grasp thereby depends upon, and is limited by, our particular
standing in the world, we are, in a profound epistemological sense,
finite. I shall argue that these theses provide crucial motivation for the
drawing of a saying/showing distinction. (My aim is not to defend them
but to see where they lead.)
One repercussion which they are often thought to have is that we can
have no knowledge unless some part of the world into which we are cast,
that part which we are to know, is given to us in some way, or (less
metaphorically) affects us in some way. This is a distinctively Kantian
thought. It occurs at the beginning of his Critique of Pure Reason*
(A19/B33) and has a primordial role to play throughout his critical
philosophy.9 Heidegger, commenting on this, recasts the thought in his
own idiosyncratic way: 'because our Dasein is finite—existing in the
midst of the essent which already is and to which our Dasein is
abandoned— . . . it [our Dasein] must of necessity receive the
essent'.10 Since what we receive thereby becomes ours, a corollary of
this thought is that what we know is thereby ours. As such, it stands in
an essential relation to us. It is 'our world'. This idea receives one of its
clearest and sharpest formulations in Schopenhauer, working very
much within the Kantian tradition. It is likewise introduced at the
beginning of his most important book, as a fundamental principle
standing in need of no proof. He writes, 'Everything that exists for
knowledge, and hence the whole of this world, is only object in relation
to the subject, perception of the perceiver, in a word, representation'.11
Anything of such fundamental importance to Schopenhauer is bound
to be relevant to this essay, given the influence which Schopenhauer
clearly exercises on Wittgenstein in the writing of the Tractatus.11
8
Immanuel Kant, Critique of Pure Reason, trans. Norman Kemp Smith
(London: Macmillan, 1933).
9
Cf. also Wittgenstein's Notebooks: 1914-1916, G. H. von Wright and
G. E. M. Anscombe (eds) and trans. G. E. M. Anscombe (Oxford: Blackwell,
1979), 74, where the idea that I am given the world is expressed explicitly
and linked in an interesting way to the will. See Colin McGinn, The Subjec-
tive View: Secondary Qualities and Indexical Thoughts (Oxford University
Press, 1983), 101-104, for one attempt to unpack the idea.
10
Martin Heidegger, Kant and the Problem of Metaphysics, trans. James
S. Churchill (Indiana University Press, 1962), 31, my emphasis.
11
Arthur Schopenhauer, The World as Will and Representation, trans.
E. F. J. Payne (New York: Dover, 1969), 3.
12
For evidence for this influence, see Allan Janik and Stephen Toulmin,
Wittgenstein's Vienna (New York: Touchstone, 1973), passim, and Nor-
man Malcolm, Ludwig Wittgenstein: A Memoir (Oxford University Press,
1958), 5.

477
A. W. Moore

Indeed a Tractarian theme is already present here, witness 5.62: 'the

world is my world'. But the important idea at which this train of
thought has arrived is that our knowledge must always somehow relate
back to us, via our reception of what we are given. Thus, still in the
Kantian tradition though this time in Fichte, we find the bold claim
that 'with respect to . . . [external] things, we know only what is
produced through our consciousness itself'.13
Now given the point of departure for this train of thought, it might
appear that a reductio ad absurdum has been affected. Indeed we do
find Fichte arguing on this basis that our knowledge is of a reality
posited by us, a conclusion which stands in direct conflict with the
original premise that our knowledge is of an independent reality not of
our own making. But there is a simple way of resisting the reductio. Let
us distinguish between what we are given and our reception of it. Our
knowledge is grounded, immediately, in the latter, which is clearly
dependent on us, but it also thereby relates, mediately, to the former,
and this need not in general be dependent on us, still less a 'posit' of
ours. It is this lack of dependence that was originally being maintained.
Does this involve a dualism between content and scheme, or between
what is given and how it is organized, or again between raw unin-
terpreted data and their interpretation? This is an urgent question,
given the severe censure to which dualisms of this kind have recently
been subjected.14
On one construal some such dualism certainly is involved—but not
on the construal which would make it objectionable. What is objec-
tionable, and what has been objected to, is the thought that within our
reception of what we are given, we can recognize a hard and fast
dichotomy between a passive element, whereby we first let the world
impinge on us, and an active element, whereby we go on to organize and
interpret what we have already thereby received. Donald Davidson has
given voice to one popular version of this thought: 'Conceptual
schemes . . . are ways of organizing experience; they are systems of
categories that give form to the data of sensation; they are points of view
from which individuals, cultures, or periods survey the passing
scene'.15 'Experience'; 'the data of sensation'; 'the passing scene': these
all already incorporate the idea of reception. By contrast, the more
fundamental dualism involved in the picture above is precisely between
our reception and what we receive (as it is in itself).
13
J. G. Fichte, The Vocation of Man, ed. and trans. Roderick M. Chisholm
(Indianapolis: Bobbs-Merrill, 1956), 74.
14
See e.g. Donald Davidson, 'On the Very Idea of a Conceptual Scheme',
reprinted in his Inquiries into Truth and Interpretation (Oxford University
Press, 1984).
15
Ibid. 183.

478
t On Saying and Showing
i

f But as well as being more fundamental, it is also more innocuous.

One can accept this dualism without thinking that the world has to be
understood apart from some way of receiving it. 'The world' here
simply plays the same kind of schematic role as the phrase 'how things
are'. Richard Rorty has complained that 'the notion of "the world" as
used in a phrase like "different conceptual schemes carve up the world
differently" must be the notion of something completely unspecified
and unspecifiable'.16 But why must it? Surely we can moot the
possibility that different conceptual schemes carve up the world
differently and thereby mean that they differ with respect to how they
treat physical objects, people, events and the like. We can even moot
the possibility that they do this without using the concepts of a physical
object or a person or an event.
Any distinction between scheme and content in the picture above is
thus an anodyne one. But it does mean that there are two facets to the
knowledge which we have. On the one hand there is that in our
knowledge which is determined by what we are given. On the other
hand there is that in our knowledge which is determined by our own
receptive capacities, which enable us to take whatever we are given. To
the extent that we can appropriately focus attention on these severally,
we can acquire different kinds of enlightenment. To acquire the first
kind of enlightenment we must exercise suitable sensitivity to what lies
outside us. To acquire the second kind of enlightenment we must be
capable of a kind of introspective, a priori reflection of our own ability
to exercise such sensitivity. In the first case we actually receive. In the
second case we are self-consciously receptive.
This is certainly the kind of distinction being sought. For one thing it
dovetails neatly with the idea that emerged from the two examples
considered earlier of a framework and its filling in, together with the
attendant idea of a subject vis-a-vis a whole. The whole in this case is
'the subject's world', that which he or she receives. But still, it might be
objected, what justification is there for relating this back to the
saying/showing distinction? In Kant, where there are these two kinds
of enlightenment, is there not also the possibility of giving voice to
both? One can say, for example, that 7+5 = 12 with no less right than
that Jesus had twelve apostles.
This objection mislocates the distinction in Kant. Certainly one can
say that 7 + 5 = 12, and much else about how the phenomenal world
must be in contrast to how it actually is (though this already presents a
contrast with the Tractatus, where propositions which are necessarily
true, that is tautologies, are deemed to say nothing (4.461)). To say
16
'The World Well Lost', reprinted in his The Consequences of Pragma-
tism (Brighton: Harvester Press, 1982), 14, his emphasis.

479
A. W. Moore '

such things is to express a priori knowledge. But it is not yet to say

anything about us or our receptive capacities as things in themselves.
The distinction between a priori knowledge and a posteriori knowledge
is not the relevant distinction. There is, in Kant, a deeper contrast
between two kinds of enlightenment, one of which does take the form of
a sort of introspective awareness of our existence as things in them-
selves—though not, of course, knowledge of ourselves as we are in
ourselves. Kant writes, 'In the synthetic original unity of apperception,
I am conscious of myself, not as I appear to myself, nor as I am in
myself, but only that I am. This representation is a thought, not an
intuition.>n Whether in Kant's view there is any legitimate way of giving
voice to this thought (this kind of enlightenment) is not clear, and it
would be rash to claim that this takes us directly back to the
saying/showing distinction. For one thing, Kant has already tried to
give voice to it in this quotation. In general when he distinguishes
between thought and knowledge, he allows for the possibility of apply-
ing concepts in expressing the former.18 Nevertheless, when set against
the enlightenment we have when intuitions and concepts are both at
play, this furnishes us with exactly the kind of contrast that can sub-
stantiate a saying/showing distinction. It is again a contrast between
knowing how things are (actually receiving) and, as a knowing subject,
having an awareness of one's own existence at the limit of what one can
know (being self-consciously receptive).
This contrast can substantiate a saying/showing distinction in the
following (Tractatus-inspired) way. To put a state of enlightenment
into words is to offer to a potential receiver (to anyone capable of
understanding the words) a model, or picture, of what one has
received. One's offering mirrors one's reception. But clearly, only
states of enlightenment which involve some reception can be treated in
this way. States of enlightenment which consist merely in being self-
consciously receptive are not susceptible to being put into words. They
indicate the limits of what can be put into words. To be in such a state of
enlightenment is to be shown something (compare 4.12-4.121).

4. A Third Example. The Tractatus Reconsidered. Its Ethical and

Mystical Aspects
We can now see how the two examples considered earlier capture the
structure of this. What the subject is shown in each case concerns how
17
Critique of Pure Reason, B157, Kant's emphasis. Cf. also his Critique of
Practical Reason, trans. Lewis White Beck (Indianapolis: Bobbs-Merril,
1956), 101.
18
See e.g. the Critique of Pure Reason, B165.

480
 On Saying and Showing

he or she stands with respect to something given, conceived as a whole.

(In the first case I am assuming that the person who assimilates the
description is the same as the person from whose point of view it is
produced.) The nature of the reception of what is given determines
both the extent of what is given (the extent of the whole) and also the
way in which the subject knows about it, and in that respect it (the
reception) imposes a framework on the whole as the subject knows it.
How the framework is filled in is a function of what is actually given,
and thus of what is received and known, and this can be put into words.
This is what can be said.
But, as remarked, the examples provide a model which is imperfect
in the following respect. In each case what the subject is shown can also
be put into words. This is because the whole is part of some larger
whole which includes the subject and which can also be given (indeed,
which he or she can be given). By contrast, for a subject to be shown
something in a full-blown Tractarian sense there must be no way of
being given the relevant larger whole. What can be shown cannot be
said.
A better example, from this point of view, would have been the
following. It amalgamates features of the other two and treats of
somebody's epistemic standing with respect to the whole of space. It
concerns a person who, like the rest of us, comes to know about the
objects in space by herself being one such and interacting with them
causally. She receives the objects in space perceptually, as an embodied
agent with various senses. What she comes to know about them can be
put into words. But her knowledge is also essentially of such a kind that
she can act on it. (This is perhaps easiest to see by reflecting on how it is
at root egocentric or perspectival, relating back to her and her own
spatial location. Fundamentally she knows such things as that the table
is immediately to the left of her.) Suppose she self- consciously reflects
on how she is able to exploit her knowledge—how, for example, know-
ing that the table is immediately to the left of her enables her to put
something on it. This will in turn entail self-conscious reflection on
how she receives the spatial objects around her: her receiving them
perceptually essentially involves her knowing how to act on them.19 But
there is no way of saying what she comes to see, or in other words what
she is shown, as she self-consciously reflects in this way. A great deal
can be said about how things are arranged in space, including how they

19
Cf. Charles Taylor, 'The Validity of Transcendental Arguments', in
Proceedings of the Aristotelian Society 79 (1979), 154; and Gareth Evans,
The Varieties of Reference (Oxford University Press, 1982), 161-162, and
'Molyneux's Question', in his Collected Papers (Oxford University Press,
1985), esp. 371 and 396-399.

481
A. W. Moore

are arranged with respect to her, and about what is required if she is
going to manipulate them in various ways. But nothing that can be put
into words will do justice to the non-inferential connection which she
perceives between the way in which spatial arrangements are presented
to her and her own agency. That she is apprised of such a connection is
part and parcel of her simply perceiving and acting. One might even
speak here of a transcendental connection between her agency and
spatiality. Certainly there is a neo-Kantian flavour to much of this.
It also has a Tractarian flavour. The connection between agency and
showing is pivotal to the ethical elements in the Tractatus, and in the
Notebooks (just as it is pivotal to much of Kant, if my way of represent-
ing Kant has been at all correct20). In accordance with the ideas which
we have been exploring the metaphysical subject in Wittgenstein's early
work receives the world in such a way as to become a limit of it (the
Tractatus, 5.632). And its will, the free exercise of which is what is
essentially good or evil (see the Notebooks pp. 79-80), determines how
it limits the world, that is it causes the world as a whole to wax or
wane—or so we should like to say (see the Tractatus, 6.43). But we
cannot strictly say this for it is what the subject is shown in adopting
different attitudes towards the world. In freely adopting these attitudes
and acting accordingly the subject is shown what is mystical; it is shown
value and meaning. For the world to wax or wane is for the world to gain
or lose meaning (the Notebooks, p. 73). A good exercise of the will
involves adopting the right attitude, the attitude of Gliick ('happiness')
(the Tractatus, 6.43 and the Notebooks, pp. 78 and 87).
However much this may puzzle us, however many questions it may
leave open, and however great may be the tension between what we are
licensed to say and what looks as if it needs saying, a way has at least
emerged of viewing the connection between the subject and the world
which makes both it and the attendant concept of showing relatively
accessible. For even if we do not know what is involved in adopting an
attitude of Gliick towards the world, that is beingglucklich (Wittgen-
stein's glosses on this as being in agreement with the world, or as doing
the will of God (the Notebooks, p. 75), do not make it any the less
schematic) we can at any rate think of it as something which the subject
does as an entity within the world, something which can legitimately be
described and which makes a sayable difference to how the world is.
The glucklich man and the unglucklich man are effably different.

20
See e.g. the 'General Note on the Transition from Rational Psychology
to Cosmology', in the Critique of Pure Reason, B428-432, and the Critique
of Practical Reason, 101. For more on the connection, see Anthony Holi-
day, 'Wittgenstein's Silence: Philosophy, Ritual and the Limits of
Language', in Language and Communication 5 (1985).

482
 On Saying and Showing

Worries about this are liable to arise on two scores and they can now be
quashed.
First there is the worry that these remarks stand in direct conflict
with the Tractarian doctrine that what can be shown cannot be said.
This worry can be quashed by reflecting on the distinction between
what is shown to somebody and what it is for that person to be shown it.
Even if the former cannot be put into words, it does not follow that the
latter cannot be. Saying what it is for somebody to be gliicklich is not the
same as saying what the world of the gliicklich man is like and does not
involve putting into words anything that is shown.21 (I do not claim to
be doing straight exegesis here. In the Notebooks, p. 78, Wittgenstein
says that there can be no describing the mark of the gliicklich life.)
The second worry is that the remarks above contradict the Trac-
tarian doctrine that the metaphysical subject cannot be found anywhere
in the world. This worry can be quashed by reflecting on the connec-
tion between the metaphysical subject and the empirical subject. There
is no need to view these as two separate entities. There is one subject
capable of being assessed from two points of view. From an external
point of view the subject is a particular human being in the world, the
empirical subject. From his or her own point of view, as recipient of the
world and set against it as a limited whole, the same subject becomes
that limiting factor outside the world to which it belongs, the
metaphysical subject (see 5.6 and following).22 A human being,
through acting in a certain way and self-consciously reflecting, can be
shown aspects of himself or herself in this latter guise. This sounds
much headier than it really is. After all, Kant has already taught us to
look upon ordinary self-consciousness as not being consciousness of
anything (any thing) in the world.23 And the perspectival differences
which arise as a result of the phenomenon of self-consciousness, though
still perhaps imperfectly understood, have become a philosophical
commonplace .24
21
For an interesting echo of this idea, see Michael Dummett, The Inter-
pretation of Frege's Philosophy (London: Duckworth, 1981), 136.
22
In David Pears, Wittgenstein (London: Fontana, 1971), 89-90, this is
related back interestingly to Wittgenstein's ethics. Cf. also H. O. Mounce,
Wittgenstein's Tractatus: An Introduction (Oxford: Blackwell, 1981), 91,
where he claims that the truth behind the solipsistic remarks is that I (the
subject) have a neighbourless point of view on the world.
23
See esp. the Paralogisms in the Critique of Pure Reason, A341-405,
and differently in the second edition, B399-432.
24
Cf. Thomas Nagel, Part I ('The Mind') of 'The Limits of Objectivity', in
The Tanner Lectures on Human Values: 1980 (Cambridge University Press,
1980), 'The Objective Self in Knowledge and Mind: Essays in Honour of
Norman Malcolm, Carl Ginet and Sydney Shoemaker (eds) (Oxford Univer-
sity Press, 1983), and The View From Nowhere (Oxford University Press,
1986), Ch. IV.

483
A. W. Moore

5. Unity in the Tractarian Distinction

What can be shown according to the Tractatus is not confined to the

ethical and mystical aspects of how the subject stands with respect to
the world. It also includes a great deal about logical form, structure and
representation. How do these tie in which each other?
In several ways. The most fundamental common factor is the idea of
the world's overall shape and form, its framework. This determines
how the facts, which constitute the world (see 1 and 2), are held
together, but it does not determine what the facts actually are. Related
to this is the idea of the world's simply existing, regardless of what it is
actually like. What can be shown is always connected not with the
world's being how it is but with its being however it is, not with what can
be said about the world but with what is involved in saying anything
about it at all (see, for example, 6.44).25 Thus what can be shown
concerns how things must be; what can be said is how they actually are.
Another common factor is meaning. It is no mere pun to say that the
ethical, mystical and logical elements in the Tractatus all concern
meaning. It highlights something fundamental about the subject's
relation to the world. Correlative with the notion of meaning is that of
understanding, and understanding lies at the root of all states of
enlightenment, whether they are instances of knowing something saya-
ble or of being shown something. This idea will be taken up in §7.
Another unifying theme linking the Tractarian remarks on showing
is unity itself. The subject, in self-consciously reflecting on the world as
his or hers and adopting an attitude towards it as a whole, must see it as
a unified whole essentially given as such. But this is possible only to the
extent that its parts are held together in unity (are united), and this
requires a unifying form—logical form. In order to represent the
world, propositions must share this form (2.18 and 4.12). If there were
no such form, that is a general form of any proposition, there would be
no such unity (theNotebooks, p. 75). This form can be given as simply,
'This is how things stand' (4.5).26
These last remarks are reminiscent of the Transcendental Deduction
in Kant's Critique of Pure Reason (A84—130 and, differently in the
second edition, Bl 16-169). In each of its versions, it traces out essen-

25
Brian McGuiness develops this idea in an extremely helpful way in his
'The Mysticism of the Tractatus'', in Philosophical Review 75 (1966).
26
Even in his later work, where Wittgenstein recoils from the importance
which he here attaches to these thoughts, he still speaks of the general
form of a proposition as a 'frame through which we look' (Philosophical
Investigations, trans. G. E. M. Anscombe (Oxford: Blackwell, 1953),
§114—see also §§134-136).

484
 On Saying and Showing

tially the same link between the unity of what can be given for con-
sciousness and the logical forms of judgments.27 Kant's concern is with
the unity that is involved in objective knowledge of the world (rather
than the mere having of experiences (see A89-90/B122)). 'A judg-
ment', he tells us, 'is nothing but the manner in which given modes of
knowledge are brought to the objective unity of apperception' (B141).
He goes on to say, 'This is what is intended by the copula "is"'. So, like
Wittgenstein, he indicates the most general form of a judgment (or
proposition) by appeal to the simple idea of asserting things to be so.
And in Kant, just as in Wittgenstein, this issues in a kind of transcen-
dental solipsism.28 For the unity involved here is the unity of appercep-
tion, the unity of what is held together in one consciousness. In the
Prolegomena29 Kant writes, 'The logical functions of all judgments are
but various modes of uniting representations in consciousness' (§22),
and in the first edition version of the Deduction he writes, 'All objects
with which we can occupy ourselves, are one and all in me, that is, are
determinations of my identical self . . . [This] is only another way of
saying that there must be a complete unity of them in one and the same
apperception. But this unity of possible consciousness also constitutes
the form of all knowledge of objects; through it the manifold is thought
as belonging to a single object' (A129). This is echoed in Wittgenstein's
claim that 'the world is my world: this is manifest in the fact that the
limits of language (of that language which alone I understand) mean
the limits of my world' (5.62, Wittgenstein's emphasis).30
We are also reminded of Frege's discussion of the unity of sentences,
which echoes Kant's claim that the copula 'is' indicates some kind of
unity.31 Frege claims, in effect, that there is no legitimate way of talking
27
See e.g. §20 of the second edition version.
28
The phrase 'transcendental solipsism' is borrowed from P. M. S.
Hacker, Insight and Illusion: Wittgenstein on Philosophy and the
Metaphysics of E x p e r i e n c e ( O x f o r d U n i v e r s i t y P r e s s , 1 9 7 2 ) , e . g . C h . I I I .
29
Prolegomena to Any Future Metaphysics, trans. Lewis White Beck
(Indianapolis: Bobbs-Merrill, 1950).
30
An alternative approach to the question of why there should be this
connection between (so-called) transcendental solipsism and logical form
is to be found in J. Hintikka, 'On Wittgenstein's "Solipsism"', in Mind 67
(1958), where he argues that Wittgenstein 'is interested only in what can
be said to be mine necessarily; for otherwise he would be doing empirical
psychology. But the only necessity there is . . . is the empty tautological
necessity of logic' (89, Hintikka's emphasis). Hintikka claims that the
metaphysical subject must therefore be identified with language; they
have the same limits.
31
See e.g. 'On Concept and Object' in Translations from the Philosophi-
cal Writings of Gottlob Frege, Peter Geach and Max Black (eds), (Oxford:
Blackwell, 1952). Cf. also in this connection Plato's Sophist, 261c6-262e2,
and Plutarch, Platonic Questions, X, 1011c.

485
A. W. Moore

about how the copula, or predicates in which it occurs, hook up with

reality. It is as if we find ourselves unintentionally talking about objects
in the world whenever we try to talk about the unifying structure which
holds them together, that which the copula hooks up with. Here it is
tempting to say precisely that we are shown how the copula hooks up
with reality though we cannot put it into words. And no doubt Witt-
genstein has Frege's problem in mind at various points in the Tractatus
(see especially the 4.12's).

6. Relativization in the Concept of Showing

Showing is grounded in the subject's reception of what he or she is

given. This suggests that there might be a deep relativization in the
concept of showing, much deeper than the relativization which was
noted in connection with the two examples in §1, which was relativiza-
tion to limited linguistic resources. For if the reception in question is
itself capable of being given to some other subject, a subject with a
radically different kind of receptive capacities, then cannot this subject
say what the original subject is shown? (We might even imagine one
subject, God say, creatively and intellectually intuiting a second sub-
ject's reception in Kant's sense, without itself being given anything or
being shown anything.32 But such a subject might also be incapable of
saying anything, in so far as saying something involves exercising
receptive capacities.)
For one subject to be able to say what another subject is shown in this
way, the two subjects would have to occupy different points of view in a
very radical sense. It is not possible to say what the Tractariangliicklich
man is shown, for example, simply by being somebody else. Any other
person receives the same world, in the same way, and in that sense
occupies the same point of view, as he does. To say what he is shown we
would have to station ourselves outside his world and thus outside our
own world, the world (compare 4.12). The impossibility of our doing
this is reflected in our ill-begotten attempts to say what he is shown. We
talk of his world's waxing, of its becoming a different world from that of
the ungliicklich man (6.43). But this makes sense only if his world is
part of, or included in, some more inclusive world to which we have
access. And it is not. There is nothing we can cite which does not
already belong to his world. So we cannot say what the gliicklich man is
shown. And this suggests that in fact, ultimately, there must be a
saying/showing distinction, such as we find in the Tractatus, which is
free of relativization. For what we cannot say cannot be said. It cannot

32
See the Critique of Pure Reason, B72.

486
 On Saying and Showing

be said in any possible language. The 'we', to use Jonathan Lear's

phrase, disappears.33
There is, however, an obvious objection to this. The 'we' disappears
only in the sense that it becomes implicit. It can easily be brought back
into view. For 'cannot be said' is simply elliptical in the above account
for 'cannot be said by us (as it were here, from this point of view)', and
'any possible language' for 'any possible language which we (or perhaps
even I) could understand' (compare 5.62). This indicates, surely, that
even the Tractarian distinction involves hidden relativization; and this
in turn reinforces the thought that some kind of relativism is a natural
concomitant of the very idea of a saying/showing distinction.
In response to this objection: making the 'we' explicit in this way
indicates genuine relativization only in so far as an alternative can be
recognized, which, in the nature of the case, it cannot. We cannot
recognize a standpoint from which that which cannot be said by us
nevertheless can be said; and we can get no proper handle on the idea of
such a standpoint's existing even though we cannot recognize it. Still,
the objection may run, even if we cannot get a proper handle on the
idea, will not self-conscious reflection at least convince us that it is
pointing to a genuine possibility and free us from what is in effect
simply a dogmatic parochialism? To this the correct response is: yes.
The saying/showing distinction is here turned on itself. For it is
through self-conscious reflection that we can be shown things, and talk
of inaccessible standpoints certainly gestures, albeit inarticulately, in
the direction of something which we can be shown. That our own point
of view is not privileged; that what we are shown may be sayable
elsewhere; that any saying/showing distinction is bound to involve
relativization; that what we take to be an absolute saying/showing
distinction we do so only because 'absolute' is itself a relative term:
these are the kinds of claims that we should like to make when giving
voice to what we are shown, though of course we cannot give voice to it,
and each of these claims, in so far as it is a genuine claim at all, is false.
What we are shown we must be content to pass over in silence (compare
7).

7. Understanding

How does understanding tie in with all of this?

The concept of understanding is extremely broad. Among the many
things that I might naturally be said to understand (or not) are:
English; some particular word or phrase; an innuendo; general

33
Op. cit., 238.

487
A. W. Moore

relativity; Kant's third Critique; why somebody is behaving in a certain

way; modern art; Beethoven's fourteenth string quartet; life; a close
friend; death. And if the solipsist really does mean something (5.62),
then what the solipsist means can be added to this list. Understanding is
one of the most general cognitive states by which we can be related to
the world. For one thing, whenever somebody knows that such and
such is the case, this will be because there is something (a situation, a
theory, a piece of information, a piece of evidence, . . .) which he or she
understands. Understanding something is always associated with hav-
ing a certain receptive capacity, and states of enlightenment of
whatever kind, including instances of being shown something, are
always grounded in some form of understanding. But our paradigm of
understanding is linguistic understanding. So too it is linguistic under-
standing which supplies our paradigms of showing.
Two facts about the Tractatus testify to this (as well as to the more
general link between showing and understanding). First, there is the
fact that the concept of showing first seriously emerges in the early 4's in
the context of a discussion of the sense of a proposition, a discussion
which focuses very much on the concept of linguistic understanding. A
proposition shows its sense, we are told; that is, it shows how things
stand if it is true, and to understand it is to know how things stand if it is
true (4.002, 4.024 and neighbouring sections).
The second relevant fact is that the solipsistic remarks, which make
their first appearance in the 5.6's somewhat out of the blue when the
discussion has hitherto been of a very impersonal, abstract, logical
character, are immediately set in the context of considerations about
linguistic understanding. As soon as we are told that what the solipsist
means is correct and that it makes itself manifest, we are given this
elucidation: 'the world is my world: this is manifest in the fact that the
limits of language (of that language which alone I understand) mean
the limits of my world' (5.62, Wittgenstein's emphasis).
This may indeed provide us with a clue as to why the solipsist is
tempted to express what he means in the way that he does (using the
pronoun 'I' rather than 'we' or 'it' or 'God'). For what he is shown
concerns his standing with respect to the world as mediated by linguis-
tic understanding; and in so far as he is inclined to see such understand-
ing as essentially and non-derivately his, he will be tempted to express
what he is shown accordingly.
Three special advantages accrue from taking linguistic understand-
ing as a prime model of being shown something (or more strictly of
what can ground being shown something). First, it throws a lot of
reciprocal light on to the philosophy of language. If we look upon
understanding in this way, then we are led to the rather radical thesis
that it is possible to show but not to state what expressions mean. This

488
 On Saying and Showing

thesis has recently received much support.34 If it is correct, then it must

have serious implications concerning the aims and aspirations which we
should have when doing the philosophy of language. In particular, it
should curtail a natural ambition to elucidate meaning by outlining
what would be involved in stating the meanings of the expressions in a
given language.35 But the thesis that it is impossible to say what any
expression means is perfectly compatible with the thesis that it is
possible to say a great deal about what it is to understand the expres-
sion, and this will be significant when it comes to deciding what we may
aspire to in practising the philosophy of language.36 It also serves to
emphasize that however mysterious and esoteric the concept of show-
ing initially threatens to be, there is no reason why instances of being
shown something should not be mundane, familiar and susceptible of
description. Understanding a language is certainly each of these things,
and this is the second reason why it is advantageous to accord it
prominence as an example of being shown something: we are reminded
of how much we can say about showing.
The third advantage is that we are provided with a helpful gloss on
what it can be like not to be shown something, in the sense of not being
sensitive to what one is capable of being shown—being prevented from
focusing attention in an appropriate way. I am not thinking in terms of
anything as straightforward as not understanding a particular language
or misunderstanding a particular word. I have in mind the case where
somebody certainly knows the meaning of a word or phrase, but
(perhaps because of a misleading analogy with some other expression)
she misconstrues its logic and thus attempts to use it in a way for which
it is unsuited. She produces nonsense, typically in the guise of
unanswerable questions, and when she self-consciously reflects on her
own linguistic understanding, she finds her attention drawn to these
'questions', which obscure what is there to be shown.
This, or something like it, is, according to the Tractatus, the general
diagnosis for (most) philosophical perplexity (4.003). And the philoso-
pher's task is obviously not to answer such 'questions' but to prevent
their arising, by reminding people of the logic of their words—of what
exactly they are entitled to say and to think—so that they are then free to
34
Cf. Michael Dummett, Frege: Philosophy of Language (London: Duck-
worth, 1973), 227, and The Interpretation of Frege's Philosophy, 129;
Gareth Evans, The Varieties of Reference, 26; Colin McGinn, 'The Struc-
ture of Content', in Thought and Object, Andrew Woodfield (ed), (Oxford
University Press, 1982), 223; and, in a somewhat different way, my 'Trans-
cendental Idealism in Wittgenstein, and Theories of Meaning', 138-139.
35
Cf. the paper cited in note 34, where I try to explore these ideas in the
light of Wittgenstein's later work.
36
Cf. again the passage from Michael Dummett cited in footnote 21.

489
A. W. Moore

focus attention on what is there to be shown though it cannot be said

(4.113—4.115 and 6.53). A valuable tool in carrying out this task will
be the injunction to look at what expressions are actuaUy used ior (for
this will highlight their logic) (6.211). Philosophy, on this conception,
is essentially an activity, whose aim is to clarify and elucidate our use of
words; there are no distinctively philosophical propositions or theses in
which it results (4.112).
Perhaps the most striking thing about this conception of philosophy
is how closely it resembles that in the later work of Wittgenstein.
Indeed the principal difference is purely terminological: Wittgenstein
later speaks of 'grammar' where here he speaks of 'logic'.37 To compare
Wittgenstein's later work with the Tractatus in these respects, and,
more specifically, to compare them in such a way that a saying/showing
distinction arises in the later work, is, I shall now suggest, deeply
instructive and seminal.38

8. The Later Work of Wittgenstein

In his later work, just as in the Tractatus, Wittgenstein urges us to view

philosophical perplexity as the result of a natural impulse to
misconstrue the grammar of our language. For example, there is an
inclination to draw false analogies between different linguistic con-
structions, the similarity of whose surface grammar hides crucial differ-
ences of function. Again, we are often tempted to interpret perfectly
good metaphors and idioms too literally and thus inappropriately to
assimilate them to utterances where the same expressions are used more
straightforwardly. The upshot is that language 'goes on holiday', good
sense gives way to nonsense, and we find ourselves puzzled and con-
fused by certain pseudo-questions which arise, posing as philosophical
problems (see the Investigations, §38). A caricature of such a 'question'
is, 'Why is it that only / can know when ouch! ?' The philosopher's task
is a therapeutic one: it is to get us to command a clear view of our
language so that we stop feeling the urge to answer such 'questions'. To
the extent that we can do this (command a clear view of our language),
we can also feel our way around inside our world-view and gain a sense
for where the limits of comprehensibility lie. That is, we can self-
consciously reflect on how we view the world, or how we receive it; we
can focus attention on our own receptive capacities without being

37
See e.g. the Philosophical Investigations, §109 and §§122-128, and
The Blue Book (Oxford: Blackwell, 1958), 26-27.
38
Cf. again the works cited in footnote 3.

490
 On Saying and Showing

distracted by ill-begotten philosophical conundrums. And, if the com-

parison with the Tractatus can indeed be pushed here, then this is
tantamount to saying that we can be shown something. An individual
who has been shown something in this way reveals this principally by
using language properly and resisting the allure of ambitious
philosophical theses with all their attendant bemusement. Such an
individual has to that extent been cured.
Here the following objection might be voiced: it is absurd to suppose
that one can reveal such a state of enlightenment simply by carrying on
in exactly the way in which one would have carried on if one had never
fallen into temptation in the first place (to switch the metaphor); for
there must surely be some further distinguishing mark which differen-
tiates one from one's unfallen but unenlightened fellows.
This objection rests on various presuppositions which are seriously
mistaken. First, it is a mistake to suppose that the individual in question
carries on in exactly the way in which he or she would have carried on
had there never been any lapse into temptation. One difference is
precisely that there has been such a lapse. The individual's reaction
against this lapse, which may be a continuing process, and which
(especially if it is) will give him or her a strong feeling of empathy for
those who still succumb to temptation, might be enough to justify talk
of a special kind of enlightenment—just as we can say that somebody
understands an argument better the livelier an appreciation he or she
has of the counterarguments to it and of how they are to be rebutted. 39
(On this point it is a notable feature of Wittgenstein's later work that in
it he always exhibits a real sense of the allure of what he is trying to
resist: the same cannot be said of those purporting to follow him.)
Secondly, there is in any case no clear warrant for thinking that not
falling into temptation entails being unenlightened. Could one not
reveal that one had been shown something just by (self-consciously)
carrying on—as it were sinlessly?
Of course, this question invites a strenuous exercise of the imagina-
tion. The metaphor of sinlessness embodies an idealization transcend-
ing anything that we ever encounter. The question should not be taken
to suggest (and is a bad question in so far as it does suggest) that in his
later work Wittgenstein distinguishes between two kinds of people:
those who need therapeutic treatment or redemption in order (once
again) to see the world aright, and those who do not, because they never

39
Cf. Jonathan Lear, op. cit., 240-241, where he writes, 'Post-neurotic
consciousness is fundamentally more complex than a healthy conscious-
ness that has never suffered disease or cure'. (Another helpful paper by
Lear which casts light on many of these issues is 'Leaving the World
Alone', in Journal of Philosophy 79 (1982).)

491
A. W. Moore

become ill or they never stray from the straight and narrow in the first
place. In fact, such a suggestion is implicit in the original objection, and
is the third and most seriously mistaken of the presuppositions on
which it rests. For underlying the objection is the following picture of
Wittgenstein's later view of philosophy: the vast majority of people
carry on exactly as they ought, but some, 'philosophers', are led astray,
and it is incumbent upon a small sub-group of these, those who have
perceived their fall into sin and who have been restored to some kind of
innocence, to expend their energies trying to redeem their fellow
philosophers. This picture is a crude travesty of what we find in
Wittgenstein's later work. The puzzles, confusions and conundrums,
not to say the deep intellectual malaise (see the Investigations, §111),
which Wittgenstein probes throughout his later work ought to be
viewed not as the foibles of some small, unfortunate group but as part of
a human predicament. It is true that Wittgenstein speaks of philosoph-
ical perplexity as the ailment of the philosopher (for example in The
Blue Book, p. 59). But we are all, in that sense, philosophers. This
aspect of the human predicament is for Wittgenstein, just as for Kant, a
natural and universal one (see the Critique of Pure Reason,
A642/B670). And there is no such thing as never succumbing to
temptation or reaching the point at last of being able to resist all
temptation. There is rather a continuing, piecemeal, non-systematic
process of trying to resist the allure of particular temptations as and
when they arise, and the philosopher's task is to try to further this
process (compare the Investigations, §127).

Now there is an irony in the Tractatus which carries over to the later
work too. Most of trie examples of philosophical nonsense which we
find in the early work do not arise as candidates for excision by the
philosophical methods outlined there. On the contrary, they actually
constitute the work, and they do so moreover to the virtual exclusion of
proper sense as construed by the Tractatus itself (see 6.54). Of course,
if it is true that philosophy is not a body of doctrine but an activity
aimed at combating nonsense (4.112), then the fact that the Tractatus
itself consists mostly of nonsense does not ipso facto prevent it from
counting as good philosophy. For it may be that uttering nonsense of a
certain kind is one way to get people to see that they themselves are
uttering nonsense (it is not hard to imagine how this could be) with the
overall effect that less nonsense is uttered. If so, then uttering nonsense
of the relevant kind would be one good way to do philosophy.40 (For
40
One could speak here of elucidatory or illuminating nonsense. Cf.
Peter Hacker, op. cit., 1-4.

492
 On Saying and Showing

that matter, it is conceivable, on this view of philosophy, that a good

way to do it would be to carry out lobotomies on people.) True, there
are parts of the Tractatus which prevent this from counting as a full
vindication of it in its own terms, for example the claim that 'the [only]
correct method in philosophy would . . . be . . . to say nothing except
what can be said' (6.53) and the claim in the Preface that if the
Tractatus has any value, it consists, in part, in the fact that (true)
thoughts are expressed in it. Nevertheless, we have here a sense of how,
even by broadly its own lights, the Tractatus can count as a successful
tissue of nonsense. Indeed, can it not eventually be turned successfully
on itself? For surely anybody who understands its purport will then
want to discard it along with other such nonsense: he or she will want to
throw away the ladder after having climbed it (6.54). The irony men-
tioned above seems not to be a deep one.
But this is too quick. The nonsense which we find in the Tractatus
cannot in fact be regarded merely as part of some clarificatory exercise.
It is a carrying out of that exercise in the wake of the kind of philosoph-
ical enlightenment which it is supposed to deliver. It is implicit in the
Tractatus that as philosophical clarification is achieved, we are shown
things. But this, ironically (and this is the original irony, which we can
now recognize as a very deep one) issues in a new impulse to utter
nonsense. We want to say what we are shown. What the Tractatus itself
consists of, for the most part, are illegitimate attempts to do just that.
So the success of the philosophical enterprise does after all bring with it
its own brand of philosophical nonsense which cannot be discarded by
any simple process of self-application.
Similarly in the later work, though in a much less marked way, the
enlightenment which is supposed to result from philosophical therapy
brings with it a new susceptibility to philosophical sickness. Self-
consciously attending to our own receptive capacities, we come to view
the world as dependent, in some hazy way, on us. We are led towards a
kind of transcendental idealism which is like the transcendental ideal-
ism of the Tractatus, except that the metaphysical subject of the earlier
work is replaced by some sort of plural descendant: there is a shift from
T to 'we'. We want to say: the limits of our language now mean the
limits of our world (compare 5.62). But this idealism, like its earlier
version, is something we are shown and there is no legitimate way of
giving voice to it. When we attempt to do so, we find ourselves making
claims which, in so far as they are intelligible at all, have to be under-
stood as straightforwardly false claims about how we (human beings, or
some particular group of human beings) stand in relation to the rest of
the world. The limits of our language (English? German? Human
discourse?) do not mean the limits of our world. They lie within it
493
A. W. Moore

awaiting empirical investigation by, say, anthropologists, sociologists

or linguists.41
The rub is that Wittgenstein himself sometimes tries in his later
work, if only via hints and suggestions, to express a kind of transcen-
dental idealism. (This is what gives this view of his later work viability.)
There are consequently times when he says (often, it is true, with some
hesitation) precisely the sorts of things which, by his own lights, ought
not to be said.42 The best examples of this are to be found in his work on
mathematics. He suggests that in some sense mathematical truth is
determined by what we believe (the Investigations, pp. 226-227).
What mathematicians would characteristically say, on the other hand,
at least when not self-consciously trying to develop a philosophy of
mathematics, is that mathematical truth is in no sense determined by
what we believe. And for Wittgenstein, what mathematicians would
characteristically say must surely be respected. It supplies, presuma-
bly, the data for any philosopher of mathematics seeking to gain a clear
view of mathematical discourse and it is not to be interfered with. The
philosopher is not, as such, in a position to take issue with any particu-
lar mathematical tenet (see the Investigations, §124).
Of course, we have to distinguish between what mathematicians
would characteristically say when doing mathematics and what they
would characteristically say otherwise, even 'when not self-consciously
trying to develop a philosophy of mathematics'. It is only the first of
these which the philosopher need in any sense treat as sacrosanct. Such
a distinction arises because everyone, the mathematician included, is
prone to confused philosophical meanderings, and there may well be a
distinctive and typical way in which mathematicians confusedly
wander off from their own discipline. Moreover, they may very well
import their confusions back into their discipline. (Intuitionists will
41
Cf. the following quotation from an unpublished early draft of John
McDowell, 'Wittgenstein on Following a Rule', in Synthese 58 (1984): 'No
doubt there is something right in finding a kind of idealism in the [Wittgen-
steinian] thought that we are guardians or trustees of meaning, not its
puppets or slaves—so that, to put it in . . . [a] dangerous way, we are
involved on the right-hand sides of statements like '"Diamonds are hard' is
true if and only if diamonds are hard". But if there is a kind of idealism here
. . . it is probably something we should do well not to try to state as a thesis
at all; but perhaps we can say that it shows itself in the relation between
language, or thought, and reality.'
42
Consider in particular some indefinite 'transcendent'-sounding uses
which he makes of the pronoun 'we', as argued convincingly by Bernard
Williams in op. cit. Two examples might be the Philosophical Investiga-
tions, §200, and, better, Philosophical Grammar, R. Rhees and trans. A. J.
P. Kenny (eds) (Oxford: Blackwell, 1974), Part II, §42.

494
 On Saying and Showing

cite use of classical reasoning as an example of this, especially, perhaps,

where false analogies are drawn between the finite and the infinite.
Think also of early work on infinitesimals.43) The lines between
mathematics and the philosophy of mathematics, or between
mathematical work which is untainted by philosophical confusion and
that which is not, or again between interfering with mathematics and
trying to combat intellectual diseases to which mathematicians are
especially prone, are not sharply defined. So it is always open to
Wittgenstein to characterize the kind of naive Platonism which fre-
quently permeates mathematical texts and which supposedly tells
against his own ruminations on mathematics, not us part of mathema-
tics, nor yet as part of a legitimate philosophy of mathematics, but as a
symptom of an extramathematical sickness indicating a need for
therapy. Indeed he writes in this vein that 'what we "are tempted to say"
. . . is . . . not philosophy; . . . it is its raw material. Thus, for
example, what a mathematician is inclined to say about the objectivity
and reality of mathematical facts, is not a philosophy of mathematics,
but something for philosophical treatment' (the Investigations, §254,
Wittgenstein's emphasis). There is nevertheless a serious problem in
that being able to distinguish between what mathematicians would say
because it is part of legitimate mathematical practice, and what they
would say because there are philosophical confusions to which they are
particularly prone, itself requires the kind of sensitivity that can be
acquired only by studying legitimate, unadulterated mathematical
practice. And even if there is no vicious circle here, the fact remains that
by his own principles Wittgenstein is entitled to describe and comment
on mathematical practice only in a way that is mathematically non-
revisionary. It is not clear that he always succeeds in confining himself
in this way.
Consider his discussion of the diagonal proof that there are more real
numbers than natural numbers (Remarks on the Foundations of
Mathematics,44 Part II, §§1—40). He expends what seems like a dis-
proportionate amount of energy in trying to hammer home that the
result of this proof, as of any other, contains no more than the proof
itself. His worry is that the usual way of putting the result suggests that
it somehow does contain more. Thus he writes, 'The dangerous, decep-
tive thing about the idea: "The real numbers cannot be arranged in a

43
Cf. Morris Kline, Mathematical Thought from Ancient to Modern Times
(Oxford University Press, 1972), 389, and Carl B. Boyer, The History of the
Calculus and its Conceptual Development (New York: Dover, 1949), 223.
44
Remarks on the Foundations of Mathematics, G. H. von Wright, R.
Rhees and G. E. M. Anscombe (eds) and trans. G. E. M. Anscombe
(Oxford: Blackwell, 1978).

495
A. W. Moore

series", or again "The set [of real numbers] . . . is not denumerable" is

that it makes the determination of a concept . . . look like a fact of
nature' (§19). An essentially related concern which he has is that the
standard ways of couching the result, and of couching correlative
results, are such as to deprive them of an anchoring in mathematical
practice. 'What', he asks, 'can the concept 'non-denumerable' be used
for? . . . [The diagonal proof] . . . may lead us to say that 2N«>K0 . . .
But if we do say it—what are we to do next? In what practice is this
proposition anchored?' (§§12 and 35). Earlier he complains, 'One
pretends to compare the 'set' of real numbers in magnitude with that of
. . . [natural] numbers' (§22).
But here one wants to reply that one pretends no such thing; one does
it (and discovers that the first set is bigger). A great deal can be said
about what the concept 'non-denumerable' can be used for and about
what follows from the proposition that 2x»>K0, how it fits into
mathematical practice, what questions it leaves open. It is perfectly
acceptable to say that there are more real numbers than natural num-
bers. Wittgenstein's worries about this are ill-founded, almost para-
noic. In striving to rid us of a confused and pernicious interpretation of
certain mathematical results, Wittgenstein casts illegitimate doubt on
the results themselves. He begins to challenge mathematical practice in
precisely the way in which he himself would insist that it cannot be
challenged. We now have, I think, an account of why this should be. He
is searching for a statement of the 'true' meaning of the mathematical
results as revealed by their proofs, the meaning we grasp as soon as we
have a really clear view of those proofs and of how they fit into the
surrounding mathematical practice, the meaning which we grasp when
we 'take a wider look round' (§6). But we grasp this meaning only in the
sense that we are shown it. It cannot be stated. As he searches for a
statement of it, Wittgenstein begins to wrench the very mathematical
discourse which is supposed to form his subject-matter. The difficulty
here, as he himself remarks elsewhere, is: to stop.45 Describing the
relevant language-games is all that he, the philosopher, can do—and all
he need do. The language-games can then speak for themselves. Again
we may quote Wittgenstein: 'Let the use of words teach you their
meaning . . . let the proof teach you what was being proved' (the
Investigations, p. 220, Wittgenstein's emphasis); 'The correct method
in philosophy . . . [is] to say nothing except what can be said' (the
Tractatus, 6.53).
There is, however, much that can be said—even by the philosopher.
Some of it is in urgent need of being said. Philosophical confusion is a
45
Zettel, G. E. M. Anscombe and G. H. von Wright (eds) and trans.
G. E. M. Anscombe (Oxford: Blackwell, 1981), §314.

496
 On Saying and Showing

universal condition of mankind and to the extent that it is deep-rooted

and exerts a damaging influence on people's lives (compare the
Investigations, § 111), then what the philosopher has to say may be of
profound practical significance. It is often mistakenly assumed that
Wittgenstein's later work, if correct, signals the end of philosophy. On
the contrary, it highlights just how much philosophical work still needs
to be done. As long as there are people reflecting on their situation and
trying to understand the world, then there will be philosophical confu-
sions needing to be teased out.46 We are always attempting to surmount
our particular, historically and culturally conditioned conceptual
apparatus, because of the allure of apparent shortcuts to a panoramic
understanding of the world: and we succeed only in mishandling the
apparatus. But even when philosophy has been put to effective work in
making us resist this allure, indeed because it has been put to effective
work, there is still a danger, though now a new danger, of philosophical
confusion. For we gain insight and are shown something which we
cannot express, though we are tempted to do so. We have to learn to say
only what can be said, to recognize the limits of our knowledge. I shall
end this essay as I began it, with a quotation from Plato:47
The purifier of the soul is conscious that his patient will receive no
benefit from the application of knowledge until he is refuted, and
from refutation learns modesty; he must be . . . made to think that
he knows only what he knows and no more.48

King's College, Cambridge

46
Cf. Kant, the Critique of Pure Reason, A298/B354-5.
47
This quotation is from the Sophist, 230c-d, trans. F. M. Cornford. The
opening quotation was from the Letters 7, 341c-d, trans. Frederick
Copleston.
48
I should like to thank Philip Turetzky for many valuable conversations
on these topics. Thanks are also due to Timothy Crane, Naomi Eilan, Ross
Harrison, Philip Percival and the editor of Philosophy for very helpful com-
ments on earlier versions of this essay.

497